Senior Engineering Manager
Islamabad · Working globallyI’m Moiz. I architect agentic AI systems and lead the teams bringing them to production, with a background in building and scaling web products.
Current work / In development
I’m building a workflow that reasons over regulatory changes, drafts updates to affected content, validates them, and creates change requests for human review.
AWS Step Functions orchestration, Bedrock tool use, and LangSmith tracing, with human review built into the workflow.
Illustrative only. A scripted simulation with made-up rules and documents, not the real product. No model is called.
The cancellation window for new subscriptions is extended from 14 days to 30 days.
The agent reads the rule, finds every page it touches, drafts the edits, and checks its own work. Nothing ships until a person approves it.
Pick a rule change and run the agent.
01 Selected work
A few projects where technical decisions
made a meaningful difference.
01 Product architecture · Technical leadership
A website relaunched in under 30 days, then grew into a community platform serving five schools. Built for a deadline, designed for what came next.
02 Full-stack engineering · Infrastructure
Real-time multiplayer escape rooms, a custom video streaming pipeline, and ownership from prototype through production.
03 Frontend systems · Developer experience
An initial marketing launch, followed by longer-term website, tooling and design system work across an 18-month engagement.
02 How I work
I’m a Senior Engineering Manager at Tintash. I stay hands-on while helping teams make good decisions, grow their craft, and take ownership of what they ship.
More about meDeadlines, security, cost, and the team who will maintain it. The right architecture starts with understanding the problem.
Architecture reviews, implementation, and production debugging. Leadership works better when it is grounded in the details.
Clear expectations, thoughtful feedback, and room to take ownership. Strong systems need people who can move them forward.
03 How I lead
Most of what I believe about leading engineers I’ve written down for my teams. Here’s the short version.
I wrote a guide for engineers early in their careers, built around six pillars. Each one pairs a core lesson with a question to sit with.
If you can’t explain the code the AI wrote, don’t commit it.
The goal in year one isn’t to know everything. It’s to learn how to learn: work at the edge of your ability, get feedback, adjust. Three years of the same work is one year repeated three times.
What did you learn this week that was at the edge of your ability?
Reasoning from first principles and an instinct for learning, over depth of knowledge.
Trade-off thinking that shows up unprompted, not tools named from memory.
Systems over heroics, ownership of outcomes, and honesty about what they don’t know.
The conversation about someone’s career and the decision about which project they join need different tools, for different audiences.
Choose the statement that matches how someone typically works. The number follows from that, so nobody can work backwards from a score they already had in mind.
Two people with the same overall score can be opposite fits for the same project. The project decides which strengths matter.
04 The workshop
Smaller tools, experiments,
and things built to learn.
Lighthouse findings, investigated against the actual source code with Claude.
Pakistan Stock Exchange data, available directly to AI assistants.
Exploring reactivity, dependency tracking, and the virtual DOM by building them.
A mock exam and study guide for Claude Certified Architect Foundations.
05 The Pixelated View
Notes on engineering, leadership, and what changes when AI becomes part of the way we work.
Have something in mind?